请选择 进入手机版 | 继续访问电脑版

基于3D DenseNet深度学习的术前CT在胸腺瘤患者中检出重症肌无力 ... ... ... ... ...

已有 94 次阅读2021-5-25 11:17 |个人分类:TET学习|系统分类:医学科学| 胸腺瘤, 重症肌无力

基于3D DenseNet深度学习的术前CT在胸腺瘤患者中检出重症肌无力
3D DenseNet Deep Learning Based Preoperative Computed Tomography for Detecting Myasthenia Gravis in Patients With Thymoma

背景:重症肌无力(MG)是胸腺瘤中最常见的副肿瘤综合症,与胸腺异常密切相关。及时发现MG的风险将有助于胸腺瘤患者的临床管理和治疗决策。在本文中,我们开发了一种基于术前计算机断层扫描(CT)的3D DenseNet深度学习(DL)模型,将其作为检测胸腺瘤患者MG的非侵入性方法。

方法:招募了230所胸腺瘤患者在医学院附属医院的大队列研究。 182名胸腺瘤患者(81名有MG,101名无MG)被用于训练和模型建立。来自另一家医院的48例病例用于外部验证。进行了3D-DenseNet-DL模型和五个放射学模型以检测胸腺瘤患者的MG。还通过集成机器学习和语义CT图像特征进行了综合分析,称为基于3D-DenseNet-DL的多模型,以建立更有效的预测模型。

结果:通过精心比较预测效果,3D-DenseNet-DL有效地识别了MG患者,并且优于其他五个放射学模型,ROC曲线下的平均面积(AUC),准确性,敏感性和特异性分别为0.734、0.724,分别为0.787和0.672。通过以下指标可以证明,基于3D-DenseNet-DL的多模型的有效性得到了进一步提高:AUC 0.766,准确性0.790,灵敏度0.739和特异性0.801。外部验证结果证实了这种基于DL的多模型的可行性,其指标分别为:AUC 0.730,准确度0.732,灵敏度0.700和特异性0.690。

解释:我们的3D-DenseNet-DL模型可根据术前CT成像有效检测胸腺瘤患者的MG。该模型可作为常规诊断标准的补充,以识别与胸腺瘤相关的MG。
关键词:计算机体层摄影术;深度学习-人工神经网络;成像-计算机断层扫描;重症肌无力;胸腺瘤。

Abstract
Background: Myasthenia gravis (MG) is the most common paraneoplastic syndromes of thymoma and closely related to thymus abnormalities. Timely detecting of the risk of MG would benefit clinical management and treatment decision for patients with thymoma. Herein, we developed a 3D DenseNet deep learning (DL) model based on preoperative computed tomography (CT) as a non-invasive method to detect MG in thymoma patients.

Methods: A large cohort of 230 thymoma patients in a hospital affiliated with a medical school were enrolled. 182 thymoma patients (81 with MG, 101 without MG) were used for training and model building. 48 cases from another hospital were used for external validation. A 3D-DenseNet-DL model and five radiomic models were performed to detect MG in thymoma patients. A comprehensive analysis by integrating machine learning and semantic CT image features, named 3D-DenseNet-DL-based multi-model, was also performed to establish a more effective prediction model.

Findings: By elaborately comparing the prediction efficacy, the 3D-DenseNet-DL effectively identified MG patients and was superior to other five radiomic models, with a mean area under ROC curve (AUC), accuracy, sensitivity, and specificity of 0.734, 0.724, 0.787, and 0.672, respectively. The effectiveness of the 3D-DenseNet-DL-based multi-model was further improved as evidenced by the following metrics: AUC 0.766, accuracy 0.790, sensitivity 0.739, and specificity 0.801. External verification results confirmed the feasibility of this DL-based multi-model with metrics: AUC 0.730, accuracy 0.732, sensitivity 0.700, and specificity 0.690, respectively.

Interpretation: Our 3D-DenseNet-DL model can effectively detect MG in patients with thymoma based on preoperative CT imaging. This model may serve as a supplement to the conventional diagnostic criteria for identifying thymoma associated MG.
Keywords: computed tomography; deep learning—artificial neural network; imaging—computed tomography; myasthenia gravis; thymoma.

  3D-DenseNet-DL和基于DL的多模型的预测指标。 使用ROC曲线下面积(AUC),ACC(准确性),SN(敏感性)和SP(特异性)量度来比较这些模型的性能。 (A)来自训练和五重交叉验证的深度学习结果的预测指标,提出的平均AUC为0.734±0.066。 (B)三种语义CT标志模型,3D-DenseNet-DL模型和综合模型(基于3D-DenseNet-DL的多模型)的比较,平均AUC分别为0.677、0.734和0.766 。 (C)在外部验证中使用3D-DenseNet-DL模型和基于3D-DenseNet-DL的多模型的值,其中DL模型使用AUC 0.704,ACC 0.690,SN 0.760和SP 0.710,而AUC 0.730,ACC 0.732,SN 最终基于3D-DenseNet-DL的多模型为0.700和SP 0.690。

Figure The prediction metrics of 3D-DenseNet-DL and DL based multi-model. The metrics Area Under ROC Curve (AUC), ACC (accuracy), SN (sensitivity), and SP (specificity) were used to compare the performance of these models. (A) The prediction metrics of the deep learning results from training and five-fold cross-validation, a mean AUC of 0.734 ± 0.066 was presented. (B) The comparison of three models of semantic CT signs model, 3D-DenseNet-DL model, and the comprehensive model (3D-DenseNet-DL based multi-model), with a mean AUC of 0.677, 0.734, and 0.766, respectively. (C) Values of 3D-DenseNet-DL model and 3D-DenseNet-DL based multi-model in external validation, with AUC 0.704, ACC 0.690, SN 0.760, and SP 0.710 for DL model, and AUC 0.730, ACC 0.732, SN 0.700, and SP 0.690 for our final 3D-DenseNet-DL based multi-model.

Liu Z, Zhu Y, Yuan Y, Yang L, Wang K, Wang M, Yang X, Wu X, Tian X, Zhang R, Shen B, Luo H, Feng H, Feng S, Ke Z. 3D DenseNet Deep Learning Based Preoperative Computed Tomography for Detecting Myasthenia Gravis in Patients With Thymoma. Front Oncol. 2021 May 5;11:631964. doi: 10.3389/fonc.2021.631964. PMID: 34026611; PMCID: PMC8132943.

路过

鸡蛋

鲜花

握手

雷人

评论 (0 个评论)

facelist

您需要登录后才可以评论 登录 | 立即注册

返回顶部